The thought is, that with an extra attacker on the ice, the chances of scoring a game tying goal are increased. But the catch of course, is that you no longer have a goalie. There’s no way the opposing team’s chances of scoring a goal are not also greatly increased. So the question is, do the benefits of having an extra player outweigh the disadvantage of not having a goalie?
Statistically, a team that pulls its goalie does increase their goal scoring chances. When even strength, the average successful shot percentage in the NHL is 5.78% and the average rebound percentage is 5.67%. With an empty net, these stats improve to 8.25% and 8.04%, meaning a team is more likely to score with an extra attacker. But it doesn’t account for how often that same empty net results in a goal for the other team. The NHL does keep stats on empty net goals, but only on the number each team has scored and the number each team has let in. Unfortunately, this doesn’t take into account the number of chances or more importantly the rate at which this is a successful move.
Many will argue it is irrelevant if a team gives up an empty net goal, as they were already trailing. This could be true if there were only one or two seconds remaining in the game, but what if the empty net goal is scored with a minute left? With the goaltender still in net, the trailing team wouldn’t have surrendered that goal, and would still have a minute left to try and tie the game. With a two goal deficit and only a minute to play, the odds of tying the game decrease dramatically, making it all but impossible.
To determine if pulling a goalie is statistically wise, we would need to know the rates that each instance is successful. To begin, we would need to determine the number of goals that a team scores even strength for every 60 minutes of play, and the rate at which they allow a goal for every 60 minutes of play. This would need to exclude all penalty situations, including when a goalie leaves the net on a delayed penalty.
This could then be compared to the rate at which the same team scores with an empty net, per 60 minutes versus how many goals they surrender with an empty net, per 60 minutes of ice time. The goal differential in each situation would then determine whether it’s a good idea or not. Basically, does the gap between goals scored and goals allowed improve with an empty net?
I suspect it doesn’t, which means that it is statistically wrong to pull your goalie, and actually hurts your chances of winning the game. But if it does, wouldn’t it make sense to pull your goalie much earlier than the final minute or two of play? The more time you play with this goal advantage, the more likely you are to capitalize on it. In fact, wouldn’t it make sense to pull your goalie anytime your team is down? The added benefit of pulling your goalie midway through the first period is that if a goal is surrendered, there is still time to comeback, while in the final minute of the game, there isn’t.
But I’m guessing that deep down, even hockey people know that pulling your goalie doesn’t really make any sense, and this is why they don’t pull their goalie every time they are down. In reality, pulling your goalie is like a Hail Mary pass in football. Sure it works the odd time, but in reality it’s merely an act of desperation, used when a team hopes the fates will take over and they’ll snatch victory from the claws of defeat.
Of course having said all this, I really can’t back it up with any statistics. Ironically, it’s just a gut feeling. Anybody want to volunteer to track this for the next four or five years?
-TheRev
Jersey Fouls!
twitter.com/TheRevBW
thesportsroundup@gmail.com
Agree? Disagree? We want to hear from you! Click below to comment
9 comments:
Hello Rev,
I am a huge fan of pulling the goalie. In fact if my team hasn't been able to score on a given night, I'd prefer to pull my goalie than send my fans to bed frustrated.
It is an interesting topic and one I think about more it seems as days go by.
About 3 goals are scored per 60 minutes in regular play. With the goalie scored, this rate goes up to 20 goals per 60 minutes (5.5 goals/60 min for the team losing and 14.5 goals/60 min for the team up). If you played 60 minutes without a goalie, which you are essentially suggesting, you would lose on average 14-6. However, you are still much more likely to score a goal.
The intuition is similar that of throwing a hail mary pass in football. The variance on scoring fluctuates wildly and the averages work against you, but you are still much more likely to score. If the goalie is pulled at a good time (like when there is an offensive faceoff), pulling the goalie is always a smart move.
i've seen goalies pulled more frequently these days with a 2 goal deficit. has his ever been successful?
I've never heard of anyone coming back from a two goal deficit by pulling the goalie...but it might have happened.
So I'm confused -- jhstew said that if you pull your goalie, you will lose 14-6, but still be more likely to score than your average 3/60 min. So you score, pulling the goalie, but still lose, as the team that's up will out score you and your empty net by 8 goals. Am I getting that right? So how is pulling the goalie a smart move? You just lost the game, by 8 goals.
You also have to know that the style of play changes dramatically in the last few minutes of play when a team is leading. The forwards are back more so when the losing team dumps and chases, it's harder to set up an attack. The extra man often times enables them to put an extra guy on the puckand creating a forecheck and an offensive plan. Without that extra guy a lot of times the losing team can't beat the trap and can't sustain any pressure in the offensive zone
I hope you saw the Bruins/Leafs game 7 last night. Boston scored two goals with less than 2 minutes remaining to tie the game....after pulling their goalie for an extra skater. It works. Rarely. But it works. An extra skater gives a team a higher chance of scoring more than keeping your goalie in. Taking your goalie out, does however, increase the chance of the other team scoring against you. Its a statistical edge, however slight it may be.
The increased probability of a goal against is essentially irrelevant, since the result of losing by one or by two or by one hundred is the same. all that matters is increasing your own probability of scoring, and extra attacker situations double it.
http://www77.homepage.villanova.edu/robert.nydick/documents/Pulling%20the%20Goalie.pdf
http://people.stat.sfu.ca/~tim/papers/goalie.pdf
Post a Comment